Re: Do Data Models Need to built on a Mathematical Concept?

"Neo" <neo55592_at_hotmail.com> wrote in message news:4b45d3ad.0305041555.283d5994_at_posting.google.com...
> > You are confusing domains and relations.
>
> I will explain my understanding so you may correct me.

Good idea.

> I see a relation as a set(person).

It is better to see a relation as a set of facts. A fact is a predicate - a true valued statement.

> Each tuple as an element(john, mary,..) of the set.

Nope. A tuple is a fact

> A tuple's attributes(gender,height) are properties of the
> element(john).

Nope, (see next comment). We do have functional dependencies however, so the value Gender might be dependent on the value of FirstName in all the tuples in a relation.

> The tuple itself represents john

If the attribute FirstName (of domain FirstNames) is a candidate key of the relation and "John" is a value of that domain, then I say it is OK to say that the value "John" represents john.
In particular there can be *many relations* in a database that have FirstName as a attribute (and maybe as a candidate key) - each relation holding different facts about john (if it has any tuples for "John")

> and no one attribute of the tuple is
> john.

Again, I say it is OK to take some domain, or set of domains, and say that an individual value of that domain(s) *is john*. This might be 'SSN', or some surrogate key or other, or in a simple database, just a FirstName.

>IMO, each tuple(row) can represents a thing in reality.

Again, a tuple is a statement about things.

Neo, have you seen and understood Hugh Darwen's quote (slightly paraphrased)

    "Nouns are to Statements as Attributes are to Tuples"

?

> A row itself in a table named Person represents a person, regardless
> of how many attributes it has.

Nouns are things. Attributes are things. Tuples and the relations that contain them are statements about things.

Be careful to not make the 'First Great Blunder'.

Regards
Paul Vernon
Business Intelligence, IBM Global Services Received on Tue May 06 2003 - 01:02:09 CEST